Image processing device and image processing method

Abstract

To provide an image processing device and an image processing method which enable a person to easily confirm manuscripts such as a receipt and a business card that easily cause false recognition.SOLUTION: A compound machine as an image processing device includes: a manuscript read-out part 13 for read out a manuscript; and an image processing part 12 including a character recognition part 12C for performing character recognition processing for the read-out manuscript, a word phrase extraction part 12X for extracting word phrases belonging to one or more terms set according to the kind of manuscripts from the word phrases obtained by the character recognition processing and displaying the extracted word phrases while being associated with the manuscript, and a read-out control part 12R for sorting the word phrases belonging to any one term, making an operation part display the manuscript associated to the word phrases and the word phrases of each of the terms associated to the manuscript in the sorted order, and receiving correction of the word phrase through the operation part.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present disclosure relates to an image processing apparatus and an image processing method that reads an original document and performs character recognition processing on the read original document. [Background technology]

[0002] Image forming devices are known that not only read documents as images, but also recognize characters from the document images. By linking the results of character recognition to documents, it becomes possible to perform, for example, character string searches or classification based on character attributes, making it easier to utilize. However, since the results of character recognition are not always correct, the reality is that important results require manual confirmation of the recognition results and correction of errors. However, manual confirmation work takes time and places a heavy burden on the person who checks. Therefore, the following have been proposed to reduce the burden of confirmation work.

[0003] When correcting one misrecognized character, image information that matches the misrecognized character is searched for, and multiple parts can be corrected at once, improving the accuracy of character correction while enabling such collective correction (see, for example, Patent Document 1). In order to improve the accuracy of character correction, the method of Patent Document 1 includes a step of selecting an arbitrary character from a character string based on an operation command from an operator, and a step of replacing the selected arbitrary character with a predetermined character according to the operation command from the operator. Then, from the character string that has been character-recognized, a character that approximately matches the selected arbitrary character is selected in order as an object of automatic correction, and the object of automatic correction is temporarily replaced with the predetermined character. Furthermore, the syntax before and after the replaced character is analyzed, and the result of the temporary replacement is determined to be valid based on the result of the analysis. In this way, only if the automatically performed character correction results in a correct syntax, the correction is determined to be valid. This improves the accuracy of character correction. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] JP 2003-223608 A Summary of the Invention [Problem to be solved by the invention]

[0005] Documents read by image processing devices include receipts and business cards. They are often smaller than the standard A4 size, and multiple documents can be read efficiently by applying a multi-crop function, for example. However, receipts often contain handwritten characters or stamps that overlap characters. Business cards often contain small characters or decorative characters. Therefore, even if character recognition is performed, there is a high possibility that the characteristics of the same characters will be determined to be dissimilar, or that different characters will be determined to have similar characteristics. There is a high possibility of misrecognition compared to general office documents printed in normal fonts. In addition, even if the document is read by an image processing device that performs automatic correction, the possibility of erroneous automatic correction cannot be denied.

[0006] The present disclosure has been made in consideration of the above circumstances, and provides a method for enabling a person to easily check documents that are prone to misrecognition, such as receipts and business cards. [Means for solving the problem]

[0007] The present disclosure provides an image processing device including a document reading unit that reads a document, a character recognition unit that performs character recognition processing on the read document, a word extraction unit that extracts words belonging to one or more items set according to the type of document from words obtained by the character recognition processing and links the extracted words to the document and displays them, and a reading control unit that sorts the words belonging to any one of the items, displays the document linked to the words in sorted order and the words of each item linked to the document on an operation unit, and accepts corrections to the words via the operation unit.

[0008] Also, from a different viewpoint, the present disclosure provides an image processing method including the steps of: a control unit of an image processing device reading a document using a document reading unit; performing character recognition processing on the read document; extracting words belonging to one or more items set according to the type of document from among the words obtained by the character recognition processing, linking the extracted words to the document and displaying them; sorting the words belonging to any one of the items, and displaying on an operation unit the document linked to the words in the sorted order and the words for each item related to the document; and accepting corrections to the displayed words. Effect of the Invention

[0009] The image processing device according to the present disclosure includes a word extraction unit that extracts words belonging to one or more items set according to the type of document and displays the extracted words linked to the document, and a reading control unit that sorts the words belonging to any one of the items, displays the document linked to the words and the words of each item linked to the document in the sorted order on an operation unit, and accepts corrections to the words via the operation unit.Therefore, by checking the document and the words of an item linked to it in sorted order, a person can easily check documents that are prone to misrecognition, such as receipts and business cards. The image processing method according to the present disclosure also achieves the same effects. [Brief description of the drawings]

[0010] [Figure 1] 1 is a block diagram showing a configuration of a multifunction peripheral shown as an example of an image forming apparatus in the present disclosure. [Diagram 2] 2 is an external perspective view of a portion corresponding to a document reading unit in the multifunction peripheral shown in FIG. [Diagram 3] 13 is an explanatory diagram showing an example of a screen that accepts settings regarding items to be extracted and items to be sorted when the type of document is a receipt; FIG. [Figure 4] 13 is an explanatory diagram showing an example of a screen for accepting settings regarding items to be extracted and items to be sorted when the type of document is a business card; FIG. [Diagram 5]11 is a first flowchart showing an example of a process related to selection of a document type and confirmation of extracted terms executed by a control unit in the first embodiment. [Figure 6] 13 is a second flowchart showing an example of a process related to selection of a document type and confirmation of extracted terms executed by the control unit in the first embodiment. [Figure 7] 4 is an explanatory diagram showing an example of a document type selection screen that the control unit causes to be displayed on the operation unit in the first embodiment. FIG. [Figure 8] 4 is an explanatory diagram showing an example of a confirmation and correction screen that a control unit causes to be displayed on an operation unit in the first embodiment; FIG. [Figure 9] FIG. 9 is an explanatory diagram showing an example of a document displayed on the confirmation and correction screen next to the example receipt shown in FIG. 8 in the second embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0011] The present disclosure will be described in further detail below with reference to the drawings. Note that the following description is illustrative in all respects and should not be construed as limiting the present disclosure. (Embodiment 1) First, an image forming apparatus according to the present disclosure will be described. Fig. 1 is a block diagram showing a configuration of a multifunction machine shown as an example of an image forming apparatus according to the present disclosure. Fig. 2 is an external perspective view of a portion of the multifunction machine shown in Fig. 1 that corresponds to a document reading unit.

[0012] As shown in FIG. 1, the multifunction device 10 includes a control unit 11, a document reading unit 13, a printing unit 14, an operation unit 15, and a storage unit 16. The control unit 11 is mainly composed of a processor and a memory as hardware resources. In addition to the processor, the control unit 11 is a circuit composed of hardware resources such as an input / output interface circuit, a timer circuit, and a communication circuit. The processor executes a control program stored in the memory to realize processing related to operation recognition and display control. In addition, jobs related to document reading and image printing, i.e., processing related to a series of image processing, are realized. In other words, the software resources and hardware resources work together to realize the function of the control unit 11 that controls the multifunction device 10.

[0013] The control unit 11 according to the present disclosure includes an operation control unit 17, a file control unit 18, a job control unit 19, and an image processing unit 12. The image processing unit 12 includes a reading control unit 12R, a printing control unit 12P, a character recognition unit 12C, and a word extraction unit 12X. The operation unit 15 includes a display unit 15D and an operation detection unit 15S.

[0014] The operation control unit 17 included in the control unit 11 is connected to the operation unit 15, recognizes the user's operation received by the operation detection unit 15S of the operation unit 15, and generates display data to be displayed on the display unit 15D of the operation unit 15. As shown in FIG. 2, the operation unit 15 is provided on the front side of the multifunction device 10. The operation unit 15 has a display device (display unit 15D) that displays the operation screen generated by the operation control unit 17. Furthermore, the operation detection device (operation detection unit 15S) that receives instructions from the user on the operation screen is included. For example, a liquid crystal display device or an OLED (Organic Light Emitting Display) can be applied as the display device, and a touch panel can be applied as the operation detection device. The operation control unit 17 communicates with the above-mentioned operation detection device and recognizes the user's operation detected by the operation detection device, that is, the instructions from the user. Also, the operation control unit 17 transmits display data to the above-mentioned display device to display it as the operation screen. The file control unit 18 included in the control unit 11 manages the data stored in the storage unit 16 and controls reading and writing.

[0015] The image processing unit 12 included in the control unit 11 generates image data of the document based on a signal of the read image output from the document reading unit 13. In the case of a scan job, the generated image data of the document is stored in the storage unit 16 or transmitted to an external device via a communication circuit (not shown). The reading control unit 12R controls the process related to reading the document. In the case of a copy job, the image data of the document is sent to the printing unit 14 and printed. In the case of a print job, image data for printing is generated based on job data received from an external device via the communication circuit, and printed by the printing unit 14. The printing control unit 12P controls the process related to printing. The character recognition unit 12C recognizes characters included in the image data of the document. The word extraction unit 12X extracts words that belong to items set according to the type of document. Then, the extracted words are linked to the document and stored at least temporarily in the storage unit 16. The character recognition unit 12C and the word extraction unit 12X will be described later.

[0016] The job control unit 19 included in the control unit 11 performs control related to the setting or execution of a job in response to an instruction from a user recognized by the operation control unit 17. Also, the job control unit 19 performs control related to the setting or execution of a job in response to an instruction from an external device as a print host via the communication circuit. For example, in the case of an instruction related to the setting of a job, the target setting is updated. In the case of an instruction to update the operation screen, the operation control unit 17 generates display data of the updated operation screen and sends it to the display device of the operation unit 15. Thereby, the operation screen is updated. When an instruction related to the execution of a job is received, the job control unit 19 controls the document reading unit 13, the printing unit 14, and the image processing unit 12 in response to the type of the received job (e.g., a scan job, a copy job, a print job), and controls the flow of image data to execute the instructed job. More specifically, the job control unit 19 recognizes the state detected by sensors (not shown in FIG. 1) arranged at various places in the document reading unit 13 and the printing unit 14. Then, it controls the operation of motors, actuators, devices, etc. (not shown in FIG. 1) arranged at the document reading unit 13 and the printing unit 14. Also, the job control unit 19 controls the image-related processing of the image processing unit 12.

[0017] The storage unit 16 includes either or both of a rewritable non-volatile memory such as a hard disk drive (HDD) or a solid state drive (SDD) and a volatile memory such as a dynamic random access memory (DRAM) for storing image data. In FIG. 1, the storage unit 16 is shown as a block separate from the control unit 11, but it may be included in the control unit 11. That is, it may be mounted on the same circuit board as the control unit 11. The storage unit 16 stores one or more items according to the type of document, and the word extraction unit 12X refers to the items according to the type of document stored in the storage unit 16 and extracts words belonging to each item. Furthermore, the storage unit 16 may store a dictionary used by the character recognition unit 12C when performing character recognition.

[0018] The document reading unit 13 reads the image of the document under the control of the job control unit 19 and converts it into a color image signal (RGB analog signal) of RGB (red, green, and blue). That is, it executes the image reading process in the copy job and the scan job. The document to be read is set by the user on the document feeder 13F or the document table 13P. The document feeder 13F transports the document one by one to a predetermined reading position. The document table 13P is made of a flat and transparent glass member. The document reading unit 13 includes an image sensor (not shown in Figs. 1 and 2). It includes a scanning mechanism (not shown in Figs. 1 and 2) for scanning the document placed on the document table 13P. The image processing unit 12 generates image data of the document based on the signal of the read image read by the image sensor. It also generates image data for printing based on the print data received from an external device via the communication circuit. The generated image data is sent to the printing unit 14 and printed, or is transmitted to an external device via the communication circuit.

[0019] The printing unit 14 feeds print sheets one by one from any of the paper feed trays 14A, 14B, and 14C that accommodate print sheets, prints an image on the print sheets, and outputs them to a discharge tray. As shown in Figures 1 and 2, the multifunction device 10 outputs printed print sheets to any of the three discharge trays 14D, 14E, and 14F.

[0020] <Extracting and verifying words from specific types of manuscripts> Next, a process will be described in which the image processing device according to the present disclosure extracts words belonging to an item from a specific type of document and allows the user to confirm the extracted words. FIG. 3 and FIG. 4 are explanatory diagrams showing an example of an extraction / sorting item selection screen that accepts user settings for items to be extracted and items to be sorted according to the type of document. The extraction / sorting item selection screen 30 shown in FIG. 3 corresponds to the case where the type of document is a receipt. The extraction / sorting item selection screen 40 shown in FIG. 4 corresponds to the case where the type of document is a business card. If the item from which words are extracted can be selected by the user, the selection of an item according to the user's intention can be accepted and words belonging to that item can be extracted, so that the processing load can be reduced by omitting the processing of the word extraction unit and the reading control unit related to unnecessary items. In addition, if the item to be sorted can be selected by the user, the document linked to the words and the words linked to the words can be displayed in the sort order of the words belonging to the item according to the user's intention, so that the user can easily confirm the extracted words.

[0021] The extraction / sort item selection screen 30 shown in FIG. 3 indicates that the items to be written on the receipt are prepared as business partner, date, amount, and recipient. The setting area 31 accepts the user's selection of which item to extract from the prepared items. Furthermore, when the scanned document is displayed for confirmation, the setting area 31 accepts the selection of which item to sort and display in sorted order. When the user touches the selection field, the operation control unit displays a circle indicating that the item has been selected. The setting area 31 indicates that the business partner, date, and amount have been selected from the prepared items. The setting area 31 also indicates that the business partner item is to be sorted. When the operation of the [Register] key 32 is recognized, the control unit 11 as the reading control unit 12R stores the setting contents of the setting area 31 in the storage unit 16. On the other hand, when the operation of the [Cancel] key 33 is recognized, the control unit 11 as the reading control unit 12R discards the setting contents of the setting area and retains the previous setting contents.

[0022] Also, the extraction / sort item selection screen 40 shown in FIG. 4 indicates that company name, name, telephone number, email address, etc. are prepared as items to be written on a business card. The setting area 41 accepts the user's selection of which item to extract from among the prepared items. Furthermore, when the scanned document is displayed for confirmation, the setting area 41 accepts the selection of which item to sort and display in sorted order. When the user touches the selection field, the operation control unit displays a circle indicating that the item has been selected. The setting area 41 indicates that company name, name, and telephone number have been selected from among the prepared items. The setting area 41 also indicates that the company name item is to be sorted. When the operation of the [Register] key 42 is recognized, the control unit 11 as the reading control unit 12R stores the setting contents of the setting area 41 in the storage unit 16. On the other hand, when the operation of the [Cancel] key 43 is recognized, the control unit 11 as the reading control unit 12R discards the setting contents of the setting area and retains the previous setting contents.

[0023] Receipts and business cards are examples of document types, but the present disclosure targets document types in which items written like these examples are expected in advance. The selection screen for items related to phrase extraction according to the document type may be displayed in response to receipt or business card being selected on document type selection screen 70 described later when scanning a document. This refers to extraction / sort item selection screen 30 shown in FIG. 3 and extraction / sort item selection screen 40 shown in FIG. 4. However, instead of displaying the extraction / sort item selection screen every time a document type is selected, it may be displayed when a specific operation (prepared to change the initial setting) on ​​operation unit 15 is received.

[0024] 5 and 6 are flowcharts showing an example of a process related to the selection of the document type and the confirmation of the extracted words and phrases executed by the control unit 11 in this embodiment. The flow of the process will be described with reference to the flowchart. It is assumed that a job related to document reading (a scan job as an example) is selected (Yes in step S11 shown in FIG. 5). When a job related to document reading is selected, the control unit 11 as the reading control unit 12R subsequently accepts the setting of the job and an instruction to start the job. In this embodiment, the control unit 11 as the reading control unit 12R accepts the setting of the document type together with the setting related to the normal job. For convenience, the flowchart in FIG. 5 shows the setting other than the document type and the instruction to start the job (step S12) separately from the setting related to the document type (step S13), but does not indicate that the setting and the instruction to start are accepted in that order. It is assumed that a document type selection screen is displayed on the operation unit 15 as one of the settings related to the job, and the selection of the type of document (document type) is accepted. In the mode in which the selection of the document type from the user is accepted, it is possible to recognize words and phrases related to items corresponding to a specific type of document designated by the user.

[0025] FIG. 7 is an explanatory diagram showing an example of a document type selection screen that the control unit 11 displays on the operation unit 15 in this embodiment. In the document type selection screen 70 shown in FIG. 7, a [Receipt] key 71 that accepts the selection of a receipt as two specific types of documents and a [Business card] key 72 that accepts the selection of a business card are displayed. In addition, a [Other than receipt / business card] key 73 for accepting the selection of other types of documents is also displayed. Note that when the operation of any key is accepted, the control unit 11 stops displaying the document type selection screen 70. When it is detected that the [Receipt] key 71 is operated on the document type selection screen 70, other settings are made, and an operation to instruct execution of a job is performed (Yes in step S15), the control unit 11 as the reading control unit 12R causes the document reading unit 13 to read documents one by one (step S21). The control unit 11 as the reading control unit 12R causes the character recognition unit 12C to perform character recognition processing on the read document. Among the recognized words and phrases, the word and phrase extraction unit 12X extracts words and phrases related to each item selected for the receipt (see FIG. 3). Then, the words and phrases extracted for each item are linked to the manuscript and stored in the storage unit 16 (step S23).

[0026] After reading, extracting words and phrases, and linking one document, the control unit 11 as the reading control unit 12R judges whether all documents have been read (step S25). If there is a document to be read next (No in step S25), the control unit 11 returns the process to the above-mentioned step S21 and reads the next document. After reading all documents (Yes in step S25), the control unit 11 sorts the read documents in the order of words and phrases related to the specified items, and displays each document in the sorted order together with the extracted words and phrases on the operation unit 15. Then, the user's confirmation and correction are accepted (step S27). In this embodiment, the sorting item related to the read receipt is "business partner" (see FIG. 3).

[0027] FIG. 8 is an explanatory diagram showing an example of a confirmation / correction screen that the control unit 11 displays on the operation unit 15 in this embodiment. As shown in FIG. 8, the control unit 11 as the reading control unit 12R displays an image of the read document (receipt) in the image display area 81 of the confirmation / correction screen 80, and displays items according to the type of document selected as item display 85. Furthermore, the extracted words for each item are displayed. In the example shown in FIG. 8, the extracted words for the business partner item, "ABCDE Co., Ltd.", are displayed in the confirmation / correction field 86, and the extracted words for the date item, "October 1, 2022," are displayed in the confirmation / correction field 87. Furthermore, the extracted words for the amount item, "10,000," are displayed in the confirmation / correction field 88. The confirmation / correction fields 86, 87, and 88 are selected and corrected by the user (step S41 shown in FIG. 6). The control unit 11 displays the [Same as previous document] key 84 to the right of the business partner confirmation / correction field 86, which is a sort item. Since the control unit 11 displays the confirmation and correction screen in sort order, the sort items are more likely to be the same words as in the previous manuscript than other items. Therefore, it is easier to find character recognition errors. It is considered that in many cases, character recognition errors in the sort items can be easily corrected by operating the [Same as Previous Manuscript] key 84 for the sort items.

[0028] On the other hand, when it is detected that the [Business Card] key 72 is operated on the above-mentioned document type selection screen 70, other settings are made, and an operation to instruct execution of a job is performed (No in step S15, then Yes in step S17), the control unit 11 as the reading control unit 12R performs the following process. That is, the document reading unit 13 reads the document one sheet at a time (step S31). The control unit 11 as the reading control unit 12R causes the character recognition unit 12C to perform character recognition processing on the read document. Among the recognized words and phrases, the word and phrase extraction unit 12X extracts words and phrases related to each item (see FIG. 4) selected for the receipt. Then, the words and phrases extracted for each item are linked to the document and stored in the storage unit 16 (step S33).

[0029] After reading, extracting words and phrases, and linking one document, the control unit 11 judges whether or not all documents have been read (step S35). If there is a document to be read next (No in step S35), the control unit 11 returns the process to the above-mentioned step S31 and reads the next document. After reading all documents (Yes in step S35), the control unit 11 sorts the read documents in order of words related to the specified items, and displays each document in the sorted order together with the extracted words and phrases on the operation unit 15. Then, the control unit 11 accepts confirmation and corrections from the user (step S27).

[0030] Also, when it is detected that the [Other than receipt / business card] key 73 is operated on the above-mentioned document type selection screen 70, other settings are made, and an operation to instruct execution of a job is performed (Yes in step S19 after No in step S17), the control unit 11 as the reading control unit 12R performs the following process. That is, the document reading unit 13 reads the document one by one and stores it in the storage unit 16 (step S37). As in the case where the document type is a receipt or business card, character recognition processing may be performed on the read document, but processing to extract words related to each item by the word extraction unit 12X and link them to the document is not performed. After reading the document one by one (loop of No in step S39) and reading all the documents (Yes in step S39), the process proceeds to step S65 in FIG. 6 described later to perform processing related to output of the read document. Details will be described later.

[0031] The process of accepting the user's confirmation and correction will be described below. For example, when the user selects and inputs corrections to the confirmation and correction fields 86, 87, and 88 of the confirmation and correction screen 80 shown in FIG. 8 (Yes in step S41 in FIG. 6), the control unit 11 as the reading control unit 12R replaces the words and phrases related to the items in the fields where corrections have been input with the ones that have been input (step S43). The corrections may be reflected not only in the manuscript where corrections have been input, but also in the same words and phrases in the same items of other manuscripts. Then, the control unit 11 returns the process to step S41 to wait for further corrections or the next operation. If further corrections are input to any of the confirmation and correction fields, the same process is repeated (Yes in step S41, a loop returning to step S41 via S43).

[0032] In the confirmation and correction screen 80 shown in FIG. 8, the control unit 11 as the reading control unit 12R accepts input of corrections in the confirmation and correction fields 86, 87, and 88, as well as operations on the [Confirm next] key 82 and the [Same as previous document] key 84.

[0033] When the control unit 11 accepts the operation of the [Same as Previous Document] key 84 (Yes in S45 after No in step S41), it performs the following process as the reading control unit 12R. That is, if there is a document that was displayed immediately before in the sorting order and confirmed / corrected, the words in the sorting items related to the currently displayed document are updated to the words in the sorting items linked to the confirmed / corrected document (step S47). Then, the process returns to the above-mentioned step S41. According to the aspect in which the [Same as Previous Document] key 84 is provided, when words in the same item are inconsistent due to erroneous recognition, it is possible to align them with the words in the immediately previous document that were correctly recognized or corrected by a simple operation.

[0034] Furthermore, when the control unit 11 accepts the operation of the [Confirm Next] key 82 on the confirmation / correction screen 80 (Yes in S49 after No in step S45), if there is an item in the field for which correction has been input, the control unit 11 as the reading control unit 12R replaces the word or phrase related to that item with the inputted one and updates the data related to that manuscript (step S51). Then, it is determined whether there is a next manuscript to be confirmed / corrected (step S61). If there is a manuscript that has not yet been confirmed / corrected (Yes in step S61), the display on the confirmation / correction screen 80 is updated to the next manuscript and the extracted word or phrase linked to that manuscript (step S63). Then, the process returns to the above-mentioned step S41 to accept operations related to confirmation and correction for the next manuscript.

[0035] As described above, on the confirmation and correction screen 80, the control unit 11 as the reading control unit 12R waits for any one of the confirmation and correction fields 86, 87, and 88, the [Confirm next] key 82, and the [Same as previous document] key 84 to be operated. In the flowchart, this corresponds to a loop that returns to S41 via No in steps S41, No in S45, and No in S49.

[0036] When all the documents have been checked / corrected in this manner (No in step S61), the control unit 11 as the reading control unit 12R performs processing related to the output of the read documents. The output destination is determined as one of the selected job types or job-related settings. The control unit 11 as the reading control unit 12R determines whether the output destination is outside or inside the multifunction device 10 (step S65). If the output destination is outside (Yes in step S65), the control unit 11 transmits data related to the read documents to the output destination device via communication (step S67) and ends the processing related to the job. On the other hand, if the output destination is inside, the control unit 11 stores data related to the read documents in the non-volatile memory of the storage unit 16 (step S69) and ends the processing related to the job. The above is the flow of processing executed by the control unit 11 in this embodiment.

[0037] (Embodiment 2) In the first embodiment, it has been described that the control unit 11 as the reading control unit 12R displays the confirmation and correction screen 80 on the operation unit 15 and displays the words of each item related to each manuscript in the sorting order. At that time, the control unit 11 as the reading control unit 12R may determine whether the words of the sorting items related to the manuscript immediately preceding the sorting order are different or similar, and display the different parts in a display mode different from the same parts. According to this mode, if there is a part that is different from the words of the same item related to the manuscript immediately preceding the sorting order, the user can easily recognize that part. FIG. 9 is an explanatory diagram showing an example of a manuscript displayed on the confirmation and correction screen after the receipt shown in the example of FIG. 8. When the receipt shown in the image display area 91 of the confirmation and correction screen 90 of FIG. 9 is compared with the receipt shown in the image display area 81 of FIG. 8, the confirmation and correction field 96 of the customer displays the erroneously recognized "ABCOE Co., Ltd.". In other words, the letter "D" is erroneously recognized as "O" instead of the receipt that should be displayed as "ABCDE Co., Ltd.", the same as the confirmation and correction field 86 of the customer of the manuscript immediately preceding the sorting order.

[0038] According to this embodiment, the control unit 11 as the reading control unit 12R judges whether or not the word in the same item in the previous manuscript is different, and therefore only the part of the character recognized as "O" is displayed with a gray background, which is different from the white background of the other parts. In this way, by displaying the parts of the sorting items that are different from the previous manuscript in a different manner from the same parts, the user can easily recognize the difference. When there is a misrecognition, the user can easily notice it, and thus the correction input can be made more reliably. The correction may be made by inputting the correct character in the confirmation correction field 96, or by operating the [same as previous manuscript] key 94. However, the correction by the [same as previous manuscript] key 94 should be applied when each item is the same as the previous manuscript. The date in confirmation and correction field 97 is also different from that in confirmation and correction field 87 in Figure 8, which is the previous manuscript. However, because date is not a sorting item, control unit 11 does not change the display mode of confirmation and correction field 97, even though there are differences. Confirmation and correction field 98 is also different from confirmation and correction field 88 in Figure 8, but the display mode is not changed. The above describes how to determine whether the sorting items are different from the previous manuscript and display them in a different manner. However, items other than the sorting items may also be displayed in a different manner if they differ from the words in the previous manuscript. However, since the confirmation and correction screen is displayed in sort order, items other than the sorting items are less likely to contain the same words as the previous manuscript compared to the sorting items.

[0039] (Embodiment 3) In the first embodiment, the type of manuscript and items corresponding to the type are prepared in advance, and the user can select an item from which to extract words. However, the type of manuscript and the items corresponding to the manuscript may be defined and edited by the user. When performing the definition and editing, several sample manuscripts may be read, and the description positions and similar word patterns common to each manuscript may be stored or machine-learned. An application to be used on an information processing device such as a PC for performing the definition and editing may be provided.

[0040] It should be understood that the present disclosure also includes any combination of the above-described aspects. In addition to the above-described embodiment, various modifications of the present disclosure are possible. These modifications should not be interpreted as not belonging to the scope of the present disclosure. The invention according to the present disclosure should include all modifications that are equivalent to the scope of the claims and that belong to the scope of the present disclosure. [Explanation of symbols]

[0041] 10: Multifunction device, 11: Control unit, 12: Image processing unit, 12C: Character recognition unit, 12P: Print control unit, 12R: Reading control unit, 12X: Word extraction unit, 13: Document reading unit, 13F: Document feeder, 13P: Document table, 14: Printing unit, 14A, 14B, 14C: Paper feed tray, 14D, 14E, 14F: Output tray, 15: Operation unit, 15D: Display unit, 15S: Operation detection unit, 16: Memory unit, 17: Operation control unit, 18: File control unit, 19: Job control unit 30,40: Extraction / sort item selection screen, 31,41: Setting area, 32,42: [Register] key, 33,43: [Cancel] key, 70: Document type selection screen, 71: [Receipt] key, 72: [Business card] key, 73: [Receipt / Other than business card] key, 80,90: Confirmation and correction screen, 81,91: Image display area, 82: [Confirm next] key, 84,94: [Same as previous document] key, 85: Item display, 86,87,88,96,97,98: Confirmation and correction field

Claims

1. a document reading unit that reads a document; a character recognition unit that performs character recognition processing on the read document; a word extraction unit that extracts words that belong to one or more items that are set according to a type of document from the words obtained by the character recognition process, and displays the extracted words in association with the document; a read control unit that sorts words belonging to any one of the items, displays on an operation unit a document linked to the words and words of each item linked to the document in the sorted order, and accepts corrections to the words and words via the operation unit; An image processing device comprising:

2. The image processing apparatus according to claim 1 , wherein the reading control unit accepts a user's selection of the items to be sorted.

3. The image processing apparatus according to claim 1 , wherein the reading control unit accepts a user's selection of an item from which words or phrases are to be extracted, among the items set according to a type of document.

4. The image processing apparatus according to claim 1 , wherein the reading control unit, when accepting correction of a word displayed on the operation unit, accepts a one-touch operation for aligning the word with a word of the same category related to a previous document in sorting order.

5. The image processing device according to claim 1, wherein the reading control unit, when displaying the words of each item relating to each manuscript in sorted order, determines whether the words are different from the words of the same item relating to the previous manuscript in sorted order, and displays the different parts in a different display manner from the same parts.

6. The image processing apparatus according to claim 1 , wherein the word extraction unit extracts words that belong to an item corresponding to a specific type of document designated by a user, and displays the extracted words in association with the document.

7. A control unit of the image processing device reading an original document using an original document reading unit; A step of performing character recognition processing on the read document; A step of extracting words and phrases belonging to one or more categories set according to the type of document from the words and phrases obtained by the character recognition process, and displaying the extracted words and phrases in association with the document; a step of sorting words belonging to any one of the items and displaying on an operation unit manuscripts linked to the words in the sorted order and the words of each item related to the manuscripts; accepting corrections to the displayed phrase; An image processing method comprising:

Images